PERFORMANCE WAS EVALUATED IN TERMS OF ARJIS' CAPABILITY TO PREDICT CRIMES AND TO EFFECTIVELY USE POLICE MANPOWER ALLOCATION. THE RESULTS OF THIS PRELIMINARY STUDY OF SAN DIEGO POLICE DEPARTMENT CRIME DATA INDICATE THAT AUTOMATIC CRIME ANALYSIS INVOLVING PATTERN RECOGNITION TECHNIQUES IS FEASIBLE. A FURTHER EFFORT SHOULD BE MADE TO DETERMINE OPTIMUM VARIABLES AND TECHNIQUES. WITH THE LIMITED VARIABLES AVAILABLE, SOME DEGREE OF ANALYSIS IS POSSIBLE USING PATTERN RECOGNITION TECHNIQUES. IT IS SHOWN THAT PERFORMANCE INCREASES AS VARIABLES RELATED TO THE PROBABILITY OF SOLUTION AND CRIME TYPE DISTRIBUTION ARE INCLUDED. IN INSTANCES OF NONSUPPRESSIBLE CRIME, THE PATTERN RECOGNITION TECHNIQUES COULD BE USED TO MAKE MORE EFFICIENT USE OF THE ASSIGNMENT OF INVESTIGATIVE PERSONNEL IN ORDER TO SOLVE THE GREATEST NUMBER OF CRIMES. IN CASES OF SUPPRESSIBLE CRIMES, PATTERN RECOGNITION'S PREDICTIVE TECHNIQUES CAN BE USED TO ASSIGN PATROL UNITS IN ORDER TO REDUCE ACTUAL OCCURRENCES OF THESE CRIMES. APPENDIXES PRESENT GEOGRAPHICAL PLOTS BY CRIME TYPE AND STATISTICS FOR SELECTED CRIME TYPES. (AUTHOR ABSTRACT MODIFIED).